From schroedr@stiefel.Stanford.EDU Wed Feb 26 00:21:58 PST 1997
X-VM-v5-Data: ([nil nil nil nil t nil nil nil nil]
	[nil "Wed" "26" "February" "1997" "00:19:57" "-0800" "Joel Schroeder" "schroedr@sccm.stanford.edu" "<199702260819.AAA03848@stiefel.Stanford.EDU>" nil "Re: SeismicJava = JOON" "^From:" nil nil "2" nil nil nil nil]
	nil)
Received: from stiefel.Stanford.EDU by spur.Stanford.EDU with ESMTP
	(1.39.111.2/16.2) id AA177015317; Wed, 26 Feb 1997 00:21:57 -0800
Return-Path: <schroedr@stiefel.Stanford.EDU>
Received: (from schroedr@localhost) by stiefel.Stanford.EDU (8.7.5/8.7.1) id AAA03848; Wed, 26 Feb 1997 00:19:57 -0800 (PST)
Message-Id: <199702260819.AAA03848@stiefel.Stanford.EDU>
X-Mailer: ELM [version 2.4 PL25]
Mime-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
From: Joel Schroeder <schroedr@sccm.Stanford.EDU>
To: martin@gpiwap1.physik.uni-karlsruhe.de, mkarren@ibm.net
Cc: matt@spur.Stanford.EDU
Subject: Re: SeismicJava = JOON
Date: Wed, 26 Feb 1997 00:19:57 -0800 (PST)

Martin,

I've put together a simple example paper using idvi.  COuld you please
check it out at:

http://sepwww.stanford.edu/sep/matt/sdi/proc/web/paper.idvi/index.html

I don't currently use activeplot, since it's just a prototype.
Try clicking on the "show" button and then click on "view" in the
"Xtp" window (and give it a little time).

Feel free to look at the corresponding makefile and paper.tex in the
directory:

~matt/sdi/proc/web

Here are a few points to keep in mind:

1. We don't use the hyperref package (yet).  I think we will avoid
using it until we find a feature it has which we can't live without.
Do you know of any such important features?

2. We use pstexpen.exp which is the same as pstexpen, but it makes the
bounding box have integer values.  For whatever reason, I was not able
to get idvi to work correctly with non-integer bounding box values.
My pstexpen.exp may have a problem, though, since (as far as I know) a
bounding box could have values of {0.20 0.10 0.40 0.40}, in which case
pstexpen.exp would produce {0 0 0 0}.  I don't know what units eps
uses, so I'm not sure if that problem would arise or not.  Do you know
if that's the case, and have you solved the problem?

3. We have, of course, not yet integrated this system into our latex2e
and makefile setup.  We want to eventually get activeplot into an idvi
form.  We're thinking of using a Perl script to translate activeplot
into a form which works with idvi.  (We're already using a
pre-processing script, so we could just throw it into that.)  Do you
have any ideas on this?

I like your idvi paper.  Movies and all!  (If I can't understand the
paper, at least I can watch it.)

Joel

> Martin: 
> 
> I send this letter one more time since I forgot to cc to your second
> account ;*)
> 
> Joel and I really like your idvi document. We have a document working
> ... almost: I describe our problems below (and Joel may give you some
> additional details).  Again, it would be nice if we do not have to
> reinvent the wheel. 
> 
> It seems that you are not including our standard reference package, but you
> do include HyperRef. Sergey claims the two conflict. Is that the reason 
> why you typed in your references? (this is not a critique .. we just want 
> to understand how you overcame the problems we encountered). 
> 
> Sorry about mailing twice.
> 
> 	Matthias
> 
> 
> Joel Schroeder writes:
>  > 
>  > 
>  > Hello Martin: 
>  > 
>  > this almost like Xmas. Joel and I have been working on Java. 
>  > We are still designing our classes but we are coming along. I am 
>  > about (next 2 weeks) to switch my research into Java. We are planning 
>  > to have a release (I guess you call that beta .. or prebeta) later 
>  > in about a month. One of our major problems is the indexing of our 
>  > Rsf (regularly sampled function), I believe. Otherwise things look 
>  > pretty good. If you want a sneak preview, come and have a look at
>  > <http://sepwww.stanford.edu/sep/matt/jag/intro/misc/notes.html> 
>  > and look for the Jag hyperlink .... otherwise you can look into 
>  > ~matt/jag on our computers. 
>  > 
>  > We have started playing around with idvi too. But Joel encountered some 
>  > problems. Our pstexpen does not seem to produce standard encapsulated 
>  > postscript (eps), nor do we get the collapsing of paragraphs. We also 
>  > have some problems with the tex imbedding of applet. Do we need to 
>  > install Hyper-Ref? Sergey claims that it conflicts with our bibliography 
>  > solution. I will ask Joel to describe to you some additional details. 
>  > 
>  > I think we would be happy to have somebody to work with us on Java. 
>  > Parallel Java sounds very exciting, but (like every good programmer 
>  > should) we have not looked at optimization at all so far (we hardly 
>  > finalize our objects). We are not even running the just-in-time compiler 
>  > by SGI. Anyhow, we would love to see a faster (and especially parallel) 
>  > version of Java combined with our work. 
>  > 
>  > I would insist, however, that our Jag code would not be "spoiled" by
>  > non-standard Java code: beyond "param" comments.  I am sure you agree
>  > that we should not trade Java's portability for speed.
>  > 
>  > I hope this information is to the point. Joel is probably leaving in a 
>  > few weeks to take a job in the computer industry. Scott Morton (SGI and
>  > Cray) and Bill Symes are interested in learning mroe about Jag. Biondo 
>  > and Jon seem to have liked the little sneak preview that we gave them 
>  > about a week ago. 
>  > 
>  > I am glad to hear from you! Thanks for that fantastic idvi example. 
>  > Please say hello to Misa and Mona. Are you coming this summer?
>  > 
>  > 	Matthias
>  > 
>  > PS: I am not sure if I mentioned it before but Pascale and I expect a 
>  > 	baby at the end of August! Yipie!
>  > 
>  > 
>  > martin@gpiwap1.physik.uni-karlsruhe.de writes:
>  >  > 
>  >  > 
>  >  > Hi Matt, Joel,
>  >  > 
>  >  > I haven't heard from either of you in a long time.
>  >  > How is your underground Java activity ?
>  >  > 
>  >  > Java is actually the reason I want to talk to you.
>  >  > 
>  >  > Would you be interested in joining forces to get 
>  >  > HCL/Goon in a PARALLEL Java environment ?
>  >  > 
>  >  > We have here a computer scientist who will do a joint
>  >  > Master's Thesis with Computer Science and Geophysics.
>  >  > 
>  >  > In particular, he will focus in writing portable
>  >  > parallel classes of some of our GOON/CLOP standard operators.
>  >  > He is very motivated and comes from a local computer science
>  >  > group that specializes in parallel Java applications
>  >  > and Java compilers.
>  >  > He will start his thesis project now and will be finished
>  >  > by the end of the year. 
>  >  > 
>  >  > So I see here an excellent opportunity to get something 
>  >  > realistically  useable going very quickly !!!
>  >  > 
>  >  > Could you let me know what you have been doing so far in Java,
>  >  > -- if any --, so that we do not duplicate here any of your efforts ?
>  >  > 
>  >  > We are also contemplating of sending him for  1-2 month to
>  >  > you if that would be of help for you. He will be writing 
>  >  > his thesis in English, so I don't see any communication problems.
>  >  > 
>  >  > 
>  >  > Please let me know as soon as possible,
>  >  > if you would like to dive into this project !
>  >  > 
>  >  > I haven't talked to Jon about it yet;
>  >  > I wanted to contact you first.
>  >  > 
>  >  > Take care,
>  >  > 
>  >  > 	Martin
>  >  > 
>  >  > 
>  >  > 
>  >  > 
>  >  > P.S. If you feel like reading a java-dvi document on the web (German):
>  >  >      http://www-gpi.physik.uni-karlsruhe.de/pub/martin/P3/idvi
>  >  > 
>  > 
>  > 
> 




